Odds are your business employs some method of operational analytics or uses another closely related method of data processing with a different name.
Whether it’s called hybrid transaction and analytics processing (HTAP), hybrid operational/analytics processing (HOAP), translytics, or continuous intelligence, what’s being described is nearly synonymous with operational analytics.
Regardless of the name, operational analytics is a business strategy of leveraging real-time information to enhance or automate decision making. It’s an attempt to replace the traditional model of forming corporate decisions around quarterly or annual reports with making responsive pivots off of data as it’s processed in the present. It’s basically turning business intelligence and analytics insights into action at the application and systems level so users can put those insights to work.
How Does Operational Analytics Work?
The key to the success of operational analytics is the timeliness and freshness of data.
Fresh data comes into an enterprise through a variety of means, whether it be analytics data gathered from mobile apps, self-submitted customer feedback forms, documentation built on a collaboration platform, or customer data entered into customer relationship management (CRM) software.
As it streams in, different enterprise departments will share data more fluidly, finding value in innovative ways. For instance, the customer support desk may cross correlate its service tickets against customer sales records and prioritize service based upon how valuable the customer is. Or product and CRM data could be combined to better target sales and marketing efforts.
Operational Analytics in Practice
In some ways, the practice of operational analytics can be said to have originated in the energy sector, which processes huge volumes of analytics and responds almost instantaneously, often with the benefit of artificial intelligence (AI).
Electricity suppliers are in a constant struggle to provide a balanced load across the energy grid, adjusting output as needed for both industrial and residential consumers.
Power consumption is gauged by the second, and as the demand goes up power plants burn hotter, boil more water, produce more steam, spin the turbines faster, and output greater amounts of electricity.
It’s a monumentally complex process that takes in data gathered across thousands of miles of infrastructure and automatically makes adjustments down to the second because even a momentary lapse in energy production is consequential.
Video game developers
Video game developers are also using operational analytics to an increasing degree, particularly as they debut actively developed products through early access programs like those on Valve’s Steam platform.
Some developers gather extensive data on player tendencies and preferences, what encounters or levels give players the most difficulty or the greatest ease, average play times, how many players actually finish the game, bugs encountered, crashes, freezes, and much more. This data is harnessed throughout the development cycle to make fixes, tweaks, buffs to weak mechanics, nerfs to overpowered ones, and so on.
This application of operational analytics has proven most valuable in competitive games, where achieving the optimal balance between characters or teams is a never-ending battle.
Online retailers have become one of the biggest and most controversial adopters of operational analytics strategies. Many retailers monitor every aspect of their customers’ behavior, serving product recommendations and ads tailored to their customers’ preferences.
These dynamic recommendations are powered by machine learning (ML)-enabled AI recommendation engines. Furthermore, even prices can be dynamic, fluctuating based on the geolocation of the customer’s IP address and potentially reflecting the in-store prices near the customer.
At the scale of a company like Amazon or Walmart, operational analytics is a necessity when it comes to inventory management as well. These companies have warehouses, distribution centers, and even trucking companies dispersed throughout all of North America.
Each day they process millions of orders, and the concentration of these orders creates the expectation that more product will need to be warehoused, more trucks will need to be supplied, and more staff will be required to pick and pack each order at the corresponding points of greatest demand.
The integration of real-time data across these organizations, even as their facilities span the continent, enable such companies to meet their rigorous supply chain demands and automatically trigger resupply orders from their partners if a shortage is anticipated.
Pitfalls of Automation
A few years ago, Amazon fully embraced its mastery of customer data, harnessing those daily analytics to produce a same-day-delivery service that would serve communities with the highest density of Amazon customers.
The expected result would bolster customer satisfaction and increase revenues while customers saved themselves a trip to the store because their intended grocery would arrive at their doorstep later that same afternoon anyway. Only through the power of data harvesting and artificial intelligence could such a strategy be successful.
However, Amazon’s AI produced an efficient same-day-delivery service map that prioritized wealthy neighborhoods and glaringly excluded poor ones. As a result, the company received backlash for what was seen as a discriminatory service map, and Amazon quietly reconsidered its approach.
Operational analytics can produce highly valuable returns for a company, but it’s important to exercise human judgment and foresight before executing what might seem like a profitable idea. Actions that customers might find irritating or even offensive have a downside that might not show up in a strict data analysis.